Comments from curator
Phenotyping data comes from the authors'
GitHub repository StevisonLab/Recombination_Hypoxia_LowTemp_in_DMEL (archived on Zenodo,
doi:10.5281/zenodo.23401297), cited in the Data Availability section of the bioRxiv preprint (
doi:10.64898/2026.10.16.759659; no journal version yet):
rawdata/byvialday.csv (marker-class counts per backcross vial and brood),
rawdata/merged.csv (the daily scoring records) and the authors' summaries in the folder output. The authors' R scripts (scripts/01_Data_Clean-up.R, recombination_updated.R, COI.R) were followed.
Design (F1 of a marker stock): about 10 virgin females of a
y cv v f stock (X-linked recessive markers
yellow,
crossveinless,
vermilion,
forked) were crossed to 5 males of each of 5 DGRP lines, and the F1 developed under one of 4 conditions: control (21% O
2, 25 °C; code
N25), hypoxia (8% O
2, 25 °C;
H25), low temperature (21% O
2, 18 °C;
N18) or both (8% O
2, 18 °C;
H18). Virgin F1 females (DGRP X /
y cv v f X) were backcrossed to males of the same DGRP line and transferred every 48 h, giving 4 broods: A (days 2-3), B (4-5), C (6-7) and D (8-9). The male progeny of each brood vial was scored for the 4 markers, which gives the crossovers of the F1 female meiosis in the intervals
y-
cv,
cv-
v and
v-
f. Values are stored under the DGRP line, with sex F (female meiosis), and the suffix _xycvvf of every phenotype name gives the cross; the condition code is in the name.
Line DGRP_042: the paper calls this line DGRP_42 and chose it from the published DGRP_042 values (chill coma recovery, paraquat survival), but gives its stock as BDSC_25193, which is DGRP_427 in the DGRP stock list (DGRP_042 is BDSC 28127). The data are stored under DGRP_042, the name used throughout the paper.
raw_1: one row per backcross vial and brood (1,028 rows, 281 backcross vials set up with F1 females of 195 F1 cross vials; these are the 281 "crosses" and 195 "backcrosses" of the paper):
-
NMales (obsolete: the denominator of the map distances): males scored for the markers. The authors removed the records of 2 of the 8 observers before any analysis (one with a triple-crossover rate 15 times the expected one, one who scored only 8 males): 2,634 of 5,531 records. byvialday.csv is the authors' file after this step and after the removal of contaminated crosses; it gives 50,209 males, as in the paper.
-
cMycv, cMcvv, cMvf: Kosambi map distance (cM) of each interval in the brood: recombinant males of the interval / males scored, converted with the Kosambi function, computed per vial and brood as in the authors' code. NA for broods with fewer than 5 males (31 rows, excluded by the authors) and when the recombinant fraction of an interval is ≥ 0.5 (16 rows, NA in the authors' code).
MapLen is the sum of the 3 distances; as in the authors' code, an NA interval is skipped in the sum. The DGRPool line means of these 4 traits equal the authors' Table 2 (output/recombination_summary_Kosambi.csv), which averages the vial x brood values over the 4 broods. The authors' models analyse the crossover and non-crossover counts per vial and brood (binomial GLMM with brood as a fixed effect and the F1 cross as a random effect).
-
BCOff: all offspring counted in the brood vial (females and males; 115,044 in total, as in the paper).
-
BCFert: "F2 fertility" of the companion preprint (Rivera-Rincón et al. 2026, bioRxiv
10.64898/2026.04.08.717251, DGRPool study 276): offspring counted within 8 days of the first count of the brood vial (cutoff of that preprint's scripts/Fertility.R) divided by the number of F1 mothers, on the same records (the 2 observers removed; with all observers the means do not match the preprint). The preprint truncates each value to an integer before averaging; it is not truncated here. These backcross data are uploaded only in this study, not in study 276.
- Obsolete design columns: BCVial_xycvvf (backcross vial), Brood_xycvvf (1-4 = A-D; the broods are rows, not separate phenotypes), F1Vial_xycvvf (F1 cross) and BCMothers_xycvvf (number of F1 mothers, 6-10).
summary: crossover interference (authors' file output/interference_summary.csv, the values of the paper's Table 3): for each line and condition, the mean (mn_) and standard deviation (sd_) over the 4 broods of the interference 1 - observed / expected double crossovers for the adjacent interval pairs
y-
cv/
cv-
v (COI12) and
cv-
v/
v-
f (COI23), and of triple crossovers over the 3 intervals (COI123). The authors compute it per brood on the counts pooled over the vials of a line and condition (broods with at least 5 males; expected counts rounded). The value of DGRP_491 under hypoxia for COI123 is -Inf in the authors' file (an expected count of 0): it is NA here. These values were reproduced from byvialday.csv with the authors' code.
Not uploaded: the 16 marker-class counts themselves (they are in byvialday.csv) and the Chillcoma/Paraquat columns of merged.csv (published DGRP values used to choose the lines).